AI: are analysts no longer needed?
The rapid progress of artificial intelligence makes us think about how the technologies and methods of analysis and decision support are transforming. Such a vision is necessary for everyone involved in the development and application of data management practices in organizations in order to understand which of them make sense to invest in. Will the familiar tools remain, or will they be completely replaced by new technologies? What place will humans occupy in these processes?
There are claims that very soon AI algorithms will cope with most analytical tasks better than people, which will leave many analysts without work. Indeed, a machine is able to very quickly "cover" a much larger volume of information than a person, take into account many factors and give a result that is most likely to be correct. The emergence of ChatGPT and other similar tools created the impression of a revolution due to the fact that the machine was able to communicate with a person in ordinary language: the disappearance of the layer between the algorithm and the decision maker forced people to talk about the uselessness of those same analysts and programmers who are now the "interface" between business leaders and the data processing environment.
What can a person do that AI cannot?
Is an AI algorithm really capable of surpassing or replacing a person? This question can be divided into two: is it technically possible, and will it bring real benefits to people, organizations, and society?
From a technical point of view, AI technologies allow us to automate some of the data analysis tasks that were previously the preserve of humans. In our article "How to use NLU and LLM in Business?" we consider the current state of penetration of these technologies into the industry and provide examples of specific successful AI use cases. Using AI to automate routine data sifting tasks allows people to focus on solving higher-level analytical problems. AI is a tool for automating certain types of work, but it is not a complete replacement for workers.
A hypothetical complete automation of decision-making processes in business cannot bring benefits. The concept of "Sustainable Growth" appeared precisely because achieving the quantitatively maximum result - growth in profits, output volume, etc. – does not guarantee getting closer to the goals that are important to humans. Decisions in business, even the most “capitalistic” ones, are always made taking into account human values. Obviously, we cannot entrust AI with making decisions about what will be better or worse from a value point of view: this will always be the prerogative of the person themselves. Trying to shift the burden of any decisions to AI carries many risks for society and business. We propose the following basic principles for the responsible use of AI in business:
- The business goal of implementing AI should be to increase revenue, not to reduce costs by saving on labor. Many entrepreneurs first think of saving money by replacing workers with AI, but it is unlikely that this will really improve economic efficiency. It is much more correct to think about how to create new digital and real products using AI, increase the volume of goods or services sold, open new markets, create new business processes, and so on. Saving on labor costs will at best reduce existing costs by a few percent, while creating new products and expanding sales can increase the company's revenue budget several times over. The potential economic effect of these measures is simply incomparable.
- Only those tasks that a person cannot perform in principle should be automated using AI: for example, finding a needle in a haystack, or looking through millions of documents in search of the necessary information. In our practice, there were examples of such tasks - no human labor would allow us to put millions of data objects in order, or compare the contents of thousands of documents to find contradictions and repetitions.
- When automating the work of workers, AI should help the worker, not control him. A person should not become part of the automated system, its extension - a physical agent, a "cog" that the system monitors and punishes him in the event of real or imaginary deviations from the ideal course of the business process.
- One of the fundamental problems of AI is its inability to critically evaluate the information used. It a priori trusts any information, no matter where it comes from. This should be kept in mind when creating AI systems that may be exposed to attempts at misuse or deception. This is one reason why it is undesirable to replace people who interact with counterparties on behalf of a company with AI - for example, clerks processing loan applications. Where a person would immediately suspect something is wrong, an AI may miss unreliable information unless special efforts were made during its creation to create tools to verify information or pass suspicious cases to a human operator.
How will AI change corporate IT?
AI tools work with text information (documents) and structured data (DB). In business applications, structured data is processed primarily. The fragmentation and poor quality of this data is an obstacle to the effective use of AI technologies, so a prerequisite for their implementation is the consolidation, validation and availability of all operational and analytical data. This task can be solved using the Enterprise Knowledge Graph tools offered by our company. AI technologies, in turn, can be used to clean and normalize the data.
Once the data problems are solved, you can begin to implement artificial intelligence technologies. We offer tools for automating communication with clients and employees (chat bot, corporate knowledge base), customer self-service, extracting structured data from text, classifying and clustering data, searching for contradictions, etc. Such solutions can improve the efficiency of employees managing business processes where such data is used.
Thus, the development of corporate IT should follow the path of preparing high-quality data, creating tools for processing it using AI, and transforming business processes using these technologies. We believe that the comprehensive development of data management practices with the use of AI as one of the tools will bring important benefits to almost any business. With a vision of technology development, we offer our customers to use data tools and practices that will remain relevant in the medium term.
